Reducing Energy Consumption in Data Centers

Laboratory for Energy Sustainable Operations, 3021 ETED Building,
Mailstop 3131, Dept. of Industrial and Systems Engineering,
Texas A&M University, College Station, TX 77843-3131.
Phone: (979) 458-2345

Project Objective:

To achieve energy-efficiency in data centers to the maximum extent possible without degrading the quality of service experienced by users.

By proactive planning capacity and managing resources in real time, energy consumption as well as greenhouse gas emissions from data centers can be significantly reduced. The methods leverage upon existing technologies such as virtualization, dynamic voltage scaling, cluster sizing and smart scheduling by meticulously combining historic and real-time information. The methods developed are integrated into a unified framework that aides decision-making at strategic, tactical or operational levels by being aware of the decision at other levels. The results are generic enough that the proposed methods can be adapted to other settings as well to reduce energy consumption.

Project Sponsor:

National Science Foundation (CMMI-0946935)

Certificate:

Data Center Operations Engineering Certificate

Project Team:

N. Gautam (Professor) website

Lewis Ntaimo (Associate Professor) website

Eduardo Perez (Former Post Doctoral Research Associate) website

Young Myoung Ko (Former Ph.D. Student) website

Julian Gallego (Former Ph.D. Student) website

Soongeol Kwon (Ph.D. Student)

Project Publications:

Gautam, N. and Ntaimo, L. "Achieving Energy-Efficiency in Data-Center Industry: A Proactive-Reactive Resource Management Framework", Proc. Of NSF CMMI Engineering Research and Innovation Conference, 2011. pdf

Gallego, J., Ko, Y.M., Polansky, R., Perez, E., Ntaimo, L. and Gautam, N. "Integrating Virtualization, Speed Scaling and Powering On/Off Servers in Data Centers for Energy Efficiency", IIE Transactions, Vol. 45, No. 10, 1114-1136, 2013. pdf

Polansky, R., Sethuraman, S. and Gautam, N. "Obtaining Optimal Thresholds for Processors with Speed-Scaling", Electronic Notes in Theoretical Computer Science, Vol. 310, 135-155, 2015. pdf

Ko, Y.M. and Gautam, N. "Critically loaded multi-server queues with abandonments, retrials, and time-varying parameters", INFORMS Journal on Computing, Vol. 5, No. 2, 2013. pdf of main paper, pdf of online supplement

Ko, Y.M. and Gautam, N. "Transient analysis of queues for peer-based multimedia content delivery", IIE Transactions, Vol. 42, No. 12, 881-896, 2010. pdf

Conference Presentations:

Gautam, N. and Rincon Mateus, C. "Adaptive Resource Allocation in Data Centers", INFORMS Annual Conference, San Diego, CA, October 2009. (Presenter: Gautam)

Gallego Arrubla, J.A., Ko, Y-M., Polansky, R., Perez, E., Ntaimo, L. and Gautam, N. "A Stochastic Programming Approach for Reducing Energy Consumption in Data Centers", INFORMS Annual Conference, Austin, TX, November 2010. (Presenter: Gallego Arrubla)

Polansky, R. and Gautam, N. "Stochastic Fluid-flow Models to Optimize Service Rate and Number of Servers", INFORMS Annual Conference, Austin, TX, November 2010. (Presenter: Polansky)

Ko, Y-M., Gautam, N. and Shakkottai, S., "On Distributed Resource Allocation for Information Management", INFORMS Annual Conference, Austin, TX, November 2010. (Presenter: Ko)

Polansky, R. and Gautam, N. "Analysis of a Stochastic Fluid-flow Model with State-dependent Output Capacities", Applied Probability Conference, Stockholm, Sweden, July 2011. (Presenter: Gautam)

Sethuraman, S., Gautam, N. and Ntaimo, L. "On Efficiently Placing Applications and Contents", INFORMS Annual Conference, Charlotte, NC, November 2011. (Presenter: Sethuraman)

Gallego Arrubla, J.A., Ko, Y-M., Polansky, R., Perez, E., Ntaimo, L. and Gautam, N. "Unified Approach to the Energy Efficiency Problem in Data Centers", INFORMS Annual Conference, Charlotte, NC, November 2011. (Presenter: Gautam)

Gautam, N. "Analysis of a System of Parallel Multi-class Non-Homogeneous Queues", INFORMS International Conference, Beijing, China, June 2012. (Presenter: Gautam)

Mohapatra, A., Mateus, C. and Gautam, N. "Controlling Energy Consumption in Data Center Servers: An MDP-based Approach", INFORMS Annual Conference, Phoenix, AZ, October 2012. (Presenter: Mohapatra)

Sethuraman, S., Gautam, N. and Ntaimo, L. "Efficient Management of Demand Loads in Cloud Computing Environments", INFORMS Annual Conference, Phoenix, AZ, October 2012. (Presenter: Sethuraman)